Privacy-Preserving Efficient Federated-Learning Model Debugging

• Published in: IEEE transactions on parallel and distributed systems Vol. 33; no. 10; pp. 2291 - 2303
• Main Authors: Li, Anran, Zhang, Lan, Wang, Junhao, Han, Feng, Li, Xiang-Yang
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.10.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Adaptation models; Algorithms; Clients; Computational modeling; Data models; data quality assessment; Debugging; Federated learning; Influence function; Laptop computers; Machine learning; Model testing; Predictive models; Privacy; Training; Training data; Visibility
• DOI: 10.1109/TPDS.2021.3137321
• ISSN: 1045-9219

Federated learning allows large amounts of mobile clients to jointly construct a global model without sending their private data to a central server. A fundamental issue in this framework is the susceptibility to the erroneous training data. This problem is especially challenging due to the invisibility of clients' local training data and training process, as well as the resource constraints. In this paper, we aim to solve this issue by introducing the first FL debugging framework, FLDebugger , for mitigating test error caused by erroneous training data. The proposed solution traces the global model's bugs (test errors), jointly through the training log and the underlying learning algorithm, back to first identify the clients and subsequently their training samples that are most responsible for the errors. In addition, we devise an influence-based participant selection strategy to fix bugs as well as to accelerate the convergence of model retraining. The performance of the identification algorithm is evaluated via extensive experiments on a real AIoT system (50 clients, including 20 edge computers, 20 laptops and 10 desktops) and in larger-scale simulated environments. The evaluation results attest to that our framework achieves accurate, privacy-preserving and efficient identification of negatively influential clients and samples, and significantly improves the model performance by fixing bugs.